Parallel bit-stream neurohardware for blind separation of sources

The use of digital stochastic computing is proposed to realize a recurrent network for blind separation of undelayed linearly superposed signals into their original components. The stochastic representation of signals allows one to design very simple digital processing elements to implement all the operations necessary for the algorithm of Herault and Jutten (1986). A first hardware implementation has been designed and the experimental results are presented.